Pyogp/Chat Logs/Daily Meeting/28 jul 2008

[9:33] Saijanai Kuhn: yo
[9:33] Enus Linden: The gangs all here!
[9:33] Enus Linden: welcome back to the work week everyone
[9:33] Enus Linden: thanks again for friday's discussions, we had some good success since...
[9:33] Tao Takashi: I only have work weekends, in the week I tend to relax ;-)
[9:34] Enus Linden: :P
[9:34] Enus Linden: so dev sandbox status...
[9:34] Enus Linden: tao provided updated docs
[9:34] Locklainn Linden: infinty not coming?
[9:34] Enus Linden: and I've had success on Mac/Win/Linux
[9:34] Enus Linden: oh
[9:34] Enus Linden: infinity isn't going to make it today
[9:34] Locklainn Linden: oh ok
[9:34] Enus Linden: i'll get status from her later
[9:35] Enus Linden: so buildout/env setup seems to not be a problem
[9:35] Locklainn Linden: new docs worked for me too
[9:35] Enus Linden: given the necessary reqs are available
[9:35] Tao Takashi: I will nevertheless try to enhance it
[9:35] Enus Linden: ty tao
[9:35] Saijanai Kuhn: and for me. just deciding which text editor to use since none of the IDEs seem useable
[9:35] Enus Linden: locklainn, you posted a jira asking for a final ZCA decision..
[9:35] Locklainn Linden: Python comes with IDLE
[9:35] Tao Takashi: like discussed earlier with Locklain: by probably providing our own package index which holds all the packages and the right versions we need
[9:35] Enus Linden: PYO-17
[9:35] Locklainn Linden: right enus
[9:36] Enus Linden: imo, we keep it
[9:36] Locklainn Linden: seems we've made 4 decisions about it already :)
[9:36] Enus Linden: and move forward
[9:36] Enus Linden: i clsoed the jira with a comment to that effect
[9:36] Locklainn Linden: how does Infinity feel about that?
[9:36] Tao Takashi: is there a way to get those new tickets automatically via email?
[9:36] Enus Linden: Infinity will accommodate our decisions i think, as long as things are stable and documented
[9:36] Locklainn Linden: ok
[9:37] Locklainn Linden: awesome
[9:37] Enus Linden: should we choose to have a oyogp.lib.core that persists, we can work on that as needed
[9:37] Locklainn Linden: with the virtual env, it won't matter what Python is installed on the stations, correct?
[9:37] Tao Takashi: I think most of the tasks we have to do will be rather repetitive
[9:37] Tao Takashi: so we can write recipes on how to solve a certain problem
[9:37] Enus Linden: i think it's a separate body of work though, to have python libs sans zca
[9:37] Tao Takashi: Locklainn: not really, it just copies that python
[9:37] Tao Takashi: and leaves out the site-packages
[9:37] Tao Takashi: it does not compile a new one
[9:37] Locklainn Linden: oh, so if its 2.3, then it uses 2.3
[9:37] Enus Linden: yes lock
[9:38] Tao Takashi: it takes the one for which you installed virtuanenv
[9:38] Enus Linden: so I'll post the location of my py2.5.2 tarball
[9:38] Enus Linden: or, use station18
[9:38] Tao Takashi: so in endeffect the one you used for starting ez_setup.py
[9:38] Tao Takashi: not using the system python is a recommended practice though
[9:38] Enus Linden: and one we'll live with for now : )
[9:39] Locklainn Linden: ok cool
[9:39] Enus Linden: so, before we status around, i'd like to point out our short term direction to you
[9:39] Enus Linden: https://wiki.secondlife.com/wiki/Pyogp/Roadmap#OGP_Beta_Milestone
[9:39] Enus Linden: the OGP beta approaches, and infinity and tess would like for pyogp to support testing the OGP enabled grid during the beta
[9:40] Enus Linden: I'd like to work toward that in the short term
[9:40] Locklainn Linden: auth and login are done, no?
[9:40] Enus Linden: and work on the buildout of a for complete Milestone1 as we go so that we can continue with that work next
[9:40] Tao Takashi: I agree on the roadmap but it would also be cool if we can discuss the direction together as a group in the future.
[9:40] Enus Linden: auth and login libs are done
[9:40] Enus Linden: tests need to be completed
[9:40] Tao Takashi: auth and login seem to be working sans bugfix for LL's AD
[9:40] Enus Linden: yes Tao, for sure
[9:40] Locklainn Linden: ok
[9:41] Tao Takashi: so establishing presence and thus making TP available will need the message stuff
[9:41] Enus Linden: it is a group effort, driven by LL's beta in the short term if that's ok
[9:41] Locklainn Linden: right
[9:41] Tao Takashi: and we also get the seedcap I think
[9:41] Tao Takashi: so place_avatar should succeed
[9:41] Tao Takashi: at least it did so sometime ago
[9:42] Locklainn Linden: right
[9:42] Enus Linden: Tao, if we don't get messaging done (a req for complete TP), for the beta, we can live with that
[9:42] Tao Takashi: well, it's sort of done in a test script so I guess we can also get it done somehow and make it cleaner later
[9:42] Tao Takashi: I at least would like to see some bot visiting my normal av
[9:42] Enus Linden: yeppers, same here
[9:43] Tao Takashi: but from what I see Locklainn is making great progress :)
[9:43] Locklainn Linden: haha hopefully it will be done soon
[9:43] Enus Linden: so, teleport really ought to 'just work' in the current libs, as long as we omit the CompleteAvatarMovement and AvatarMovementComplete messages, yes?
[9:43] Locklainn Linden: there are some things that need to be done, that I see
[9:43] Tao Takashi: maybe I will look into the loop stuff and make some demo loop with the code from the example which we can replace with Locklainn's stuff then
[9:44] Tao Takashi: for now I guess we can live with one main loop without and threads, coroutines etc.
[9:44] Locklainn Linden: well,
[9:44] Locklainn Linden: I don't think the current code has anything for udp
[9:44] Locklainn Linden: enus, your old code does
[9:44] Tao Takashi: but as said on IRC, I would also put that loop into some example app and not the lib.
[9:44] Saijanai Kuhn: the presence code sends COmpeteAvatarMovement and the like via UDP. Is there a reason why you want to avoid that?
[9:44] Locklainn Linden: so maybe we could reference that
[9:44] Tao Takashi: I mean Enus' code
[9:45] Locklainn Linden: oh ok
[9:45] Tao Takashi: or Saijanai's/Enus' code
[9:45] Locklainn Linden: right
[9:45] Locklainn Linden: Sai, what do you mean avoid it?
[9:45] Enus Linden: Sai, I am not avoiding it, I am just not making it a requirement of testing teleport in support of the OGP Beta (unless it's unavavoidable)
[9:45] Saijanai Kuhn: those are hand assembled packets but they're pretty trivial One or two lines of code each
[9:45] Tao Takashi: so I might simply write some example app which uses the lib to the point we have now and add the UDP code to it
[9:46] Enus Linden: to test teleport, i am happy to 1. login and rez on a sim 2. tp to a different sim and get the seed cap from there
[9:46] Tao Takashi: might be interesting to see if this works with opensim
[9:46] Saijanai Kuhn: From what Tess was saing, you maintain presence by doing the long poll rather than ACKing incoming UDP packets
[9:46] Enus Linden: yes Tao, I'd like for that to be a test case we show
[9:46] Locklainn Linden: you maintain AD presence Sai
[9:46] Saijanai Kuhn: oh, OK
[9:46] Enus Linden: that is the next big component to talk about imo Sai
[9:46] Tao Takashi: so what is the state of the docs?
[9:47] Saijanai Kuhn: still need the ack for Sim presence. Got it
[9:47] Locklainn Linden: but I think you still need to maintain sim presence
[9:47] Tao Takashi: yesterday I was poking around opensim to find out what might be going on
[9:47] Locklainn Linden: we should probably make sure of that
[9:47] Tao Takashi: maybe not the best way ;-)
[9:47] Han Bookmite: is that your boat?
[9:47] Enus Linden: which dosc specifically tao?
[9:47] Tao Takashi: teleport
[9:47] Enus Linden: i need to check on that this morning tao
[9:47] Enus Linden: leyla needs to post updated login docs
[9:48] Tao Takashi: cool, thanks
[9:48] Tao Takashi: so for doing TP I basically send another place_avatar?
[9:48] Enus Linden: and i believe the names of the rez_avatar caps are changing from place_avatar to rez_avatar/place
[9:48] Enus Linden: lock, is that right?
[9:48] Tao Takashi: and the AD knows that I am still in sim A and does everything needed to move you to sim B?
[9:48] Enus Linden: we still support place_avatar currently tho
[9:48] Locklainn Linden: I believe that is what tess said, but I haven't looked in awhile
[9:48] Tao Takashi: why do we have such strange caps names now?
[9:48] Enus Linden: correct Tao
[9:48] Enus Linden: we ought to ask tess
[9:49] Enus Linden: something about namespace
[9:49] Enus Linden: i just wrote it down to confirm with tess
[9:49] Enus Linden: but i believe the family is
[9:49] Tao Takashi: "request" is what opensim expects
[9:49] Enus Linden: rez_avatar/place
[9:49] Enus Linden: rez_avatar/request
[9:49] Enus Linden: rez_avatar/derez
[9:49] Locklainn Linden: [1]
[9:49] Enus Linden: rez_avatar/rez
[9:49] Enus Linden: yep locklainn
[9:49] Locklainn Linden: it says rez_avatar/place
[9:50] Enus Linden: but the flow pic has the old naming
[9:50] Locklainn Linden: right
[9:50] Tao Takashi: I cannot really make sense of that big flow diagram there ;-)
[9:50] Enus Linden: all we care about is the client column :)
[9:50] Tao Takashi: I also care about the agent domain
[9:50] Enus Linden: so, that said, anyone game for mocking up a sample teleport?
[9:50] Locklainn Linden: Enus, he has been working on making his own AD
[9:51] Enus Linden: i know i know
[9:51] Enus Linden: : )
[9:51] Locklainn Linden: ok, just making sure :)
[9:51] Tao Takashi: the LL AD changes too much ;-)
[9:51] Enus Linden: i agree tao
[9:51] Tao Takashi: well, and I cannot debug there myself
[9:51] Enus Linden: and the docs don't precede the changes
[9:51] Enus Linden: or even follow the changes properly
[9:51] Saijanai Kuhn: there are docs?
[9:51] Enus Linden: Zero is on the hook for an updated ogp draft
[9:52] Tao Takashi: well, it shouldn't be too hard to do another place_avatar after the first one
[9:52] Tao Takashi: but can I reuse the cap or do I need to get a new one?
[9:52] Locklainn Linden: which cap
[9:52] Enus Linden: we also need this: docs supporting executin interop tests (and a more complete test suite)
[9:52] Locklainn Linden: AD cap or sim cap?
[9:52] Tao Takashi: place_avatar
[9:52] Saijanai Kuhn: where in the beta client is the code, do you know?
[9:52] Enus Linden: i'll catch up with infinity about pyogp.interop
[9:52] Locklainn Linden: hmm
[9:52] Tao Takashi: or rez_avatar/place as it's called now
[9:52] Locklainn Linden: good question
[9:53] Locklainn Linden: my guess is you can reuse it
[9:53] Locklainn Linden: because its a cap to the AD
[9:53] Tao Takashi: I hope the spec also says what happens should it be timed out
[9:53] Tao Takashi: probable 403
[9:53] Tao Takashi: or 404?
[9:53] Enus Linden: Leyla should provide that to use soon tao
[9:53] Tao Takashi: ok
[9:53] Enus Linden: so, seed cap: which one are we talking abuot?
[9:54] Tao Takashi: what seed cap? :)
[9:54] Locklainn Linden: he was talking about place_avatar cap
[9:54] Enus Linden: sorry
[9:54] Locklainn Linden: I thought seed cap as well
[9:54] Enus Linden: re use i believe
[9:54] Tao Takashi: well, I hope I can reuse the seed cap :)
[9:54] Tao Takashi: otherwise it would be strange ;-)
[9:54] Enus Linden: we can test it though :)
[9:55] Enus Linden: so, any statuses anyone wants to share?
[9:55] Tao Takashi: I should write some book: OGP - annotated
[9:55] Locklainn Linden: builder done, reader probably 75%
[9:55] Saijanai Kuhn: should be one seed cap per region
[9:55] Enus Linden: w00t lock!
[9:55] Enus Linden: expect to finish that when?
[9:55] Locklainn Linden: I'm hoping today
[9:56] Enus Linden: cool. might i be able to get you to help tess and me write some interop tests this week?
[9:56] Saijanai Kuhn: in fact, I think libsl checks to see if you hhave encountered a seedcap before and resuses it rather than makign a new entry
[9:56] Locklainn Linden: sure Enus
[9:56] Tao Takashi: I will have a look at adding some exceptions
[9:56] Enus Linden: cool. i've created some jiras and will start doling them out lock
[9:56] Tao Takashi: should time permit it, we also have a deadline here
[9:57] Locklainn Linden: awesome. When do you want to discuss the details of this?
[9:57] Tao Takashi: and add some tests for these exceptions
[9:57] Enus Linden: nice tao
[9:57] Enus Linden: can you create a jira for that?
[9:57] Tao Takashi: I will make some proposal in code, I guess we don't need that much
[9:57] Enus Linden: and add some line to the roadmap as a starter for trackin git?
[9:58] Tao Takashi: it really feels more like some corp project than an FLOSS project ;-)
[9:58] Enus Linden: lock, i'll sort out tasks today and let you pick what you want...
[9:58] Locklainn Linden: awesome
[9:58] Tao Takashi: *creating jira*
[9:58] Enus Linden: Tao, let's talk about how to change that : )
[9:59] Tao Takashi: let me first find the button for "new ticket" ;-)
[9:59] Tao Takashi: jira needs some usability overhaul
[9:59] Enus Linden: i need a way to track work, jira seems like the right way to me
[9:59] Enus Linden: agreed on usability
[9:59] Enus Linden: huddle ajourned....
[9:59] Enus Linden: transcript forthcoming
[9:59] Enus Linden: so tao
[10:00] Enus Linden: corp vs floss
[10:00] Enus Linden: it's a marraige b/n the 2 right now in some ways
[10:00] Locklainn Linden: what is the difference
[10:00] Enus Linden: I will try to drive the LL needs
[10:00] Enus Linden: and work it as an os project too
[10:00] Enus Linden: but could use some help per the later
[10:01] Enus Linden: diff... hmm. LL has deliverables and dates and company driven needs
[10:01] Tao Takashi: ok, jira is there
[10:01] Enus Linden: ty tao
[10:01] Tao Takashi: yes, you have deadlines ;-)
[10:02] Enus Linden: admittedly, the beta deadline wrt pyogp was a surprise to me
[10:02] Tao Takashi: and PR issues I think
[10:02] Enus Linden: i did sign up for a deliverable
[10:02] Tao Takashi: because nobody else would have announced a beta ;-)
[10:02] Enus Linden: to have pyogp running tests against OGP enabled grids by the end of Q3 (sept)
[10:03] Tao Takashi: well, not PR issues but PR influence
[10:03] Locklainn Linden: is that end of sept, or before enu?
[10:03] Enus Linden: other than that, infinity's urgency was new to me last week.
[10:03] Enus Linden: lock, that's what i personally signed up for
[10:03] Enus Linden: things have taken a turn
[10:03] Enus Linden: :D
[10:03] Locklainn Linden: haha yea, I feel teh same
[10:03] Saijanai Kuhn: eh the announcement was same day as google announcement. PIggybacking, I think
[10:04] Locklainn Linden: all of a sudden we had deadlines haha
[10:04] Enus Linden: so pyogp is desired as having tests to support the beta functionality
[10:04] Tao Takashi: I still call it dev sandbox functionality ;-)
[10:04] Enus Linden: so, I'll see about meeting that goal
[10:04] Tao Takashi: beta would imply OGP would be finished and we just polish things ;-)
[10:04] Tao Takashi: so I see this as PR driven
[10:04] Saijanai Kuhn: sure, but by timing it when they did. They got honorable mention in all the articles about Google Lively
[10:04] Tao Takashi: but anyway, now we have that beta and you have deadlines ;-)
[10:05] Enus Linden: nomenclature aside (thanks for having better english than me tao), yes, PR is driving things to a degree
[10:05] Bartholomew Kleiber: as it always is :-/
[10:05] Ricken Flow: smiles
[10:05] Enus Linden: so, I'll deliver LL driven deliverable requests to the group
[10:05] Enus Linden: and we can try and meet them
[10:05] Tao Takashi: so we just need to communicate these things
[10:05] Enus Linden: LL resources can drive that
[10:05] Enus Linden: Tao, what I'd like to work with you, is the higher level project roadmap
[10:06] Enus Linden: and dev methods and such
[10:06] Tao Takashi: and it would be good if not some pyogp deadline or announcement will happen without the group's knowledge
[10:06] Enus Linden: i'm with you there
[10:06] Enus Linden: so, it is now known :P
[10:06] Enus Linden: and I'd like to coordinate the future work starting as soon as we can huddle on that
[10:06] Bartholomew Kleiber: so what is the exact deadline?
[10:07] Enus Linden: vague bartholomew
[10:07] Enus Linden: beta starts july 31
[10:07] Bartholomew Kleiber: ok, who will be nailing it?
[10:07] Enus Linden: i'd like to be able to point the beta group to a wiki page saying: pyogp supports testing of the OGP implementation in these ways
[10:07] Enus Linden: hence https://wiki.secondlife.com/wiki/Pyogp/Roadmap#OGP_Beta_Milestone
[10:08] Bartholomew Kleiber: ok
[10:08] Multi Gadget: v2.0.3b by Timeless Prototype, '/44 info'
[10:08] Saijanai Kuhn: we also need a standalone easy install just for the tester
[10:08] Enus Linden: nailing it?
[10:08] Enus Linden: right sai
[10:08] Bartholomew Kleiber: like defining the exact date
[10:09] Tao Takashi: ok, so what did I miss? :)
[10:09] Enus Linden: nada tao
[10:09] Enus Linden: i won't commit to date yet bartholomew, as i need to talk to infinity about pyogp.interop
[10:09] Bartholomew Kleiber: I see, ok
[10:09] Tao Takashi: date for what?
[10:09] Enus Linden: sai raises a good point, we need to be able to run interop tests for the beta
[10:09] Bartholomew Kleiber: for the deadline
[10:10] Tao Takashi: ok, but does this need to be some standalone app?
[10:10] Tao Takashi: I mean, it's from devs for devs ;-)
[10:10] Enus Linden: i'm not willing to set a hard deadline for the pyogp deliverables. next week sounds nice though... i beliecve that is attainable
[10:11] Tao Takashi: well, we should think about releasing something as soon as some milestone is finished.. More thinking about the lib here though than the tests
[10:11] Locklainn Linden: on that note, have we decided how tests are going to be configured?
[10:11] Enus Linden: tao, if we can point people to the dev sandbox page and the client's tests, and the, point people to interop and it's tests (to be further deved this week) great
[10:11] Locklainn Linden: config settings for 1 grid may not work for another
[10:11] Enus Linden: that is yet to be done i believe lock
[10:11] Tao Takashi: yes, this probably should be done with Infinity
[10:12] Locklainn Linden: I see
[10:12] Tao Takashi: not sure how you best pass config to unit tests
[10:12] Enus Linden: infinity was planning to propose a method
[10:12] Tao Takashi: who knows maybe in the end unit tests are not the right tool
[10:12] Enus Linden: of passing in data from xml files
[10:12] Enus Linden: unit tests are adequate for now imo
[10:13] Tao Takashi: well, I leave that for you guys :)
[10:13] Locklainn Linden: yea, for now, but they will al lbe configured for our grid
[10:13] Saijanai Kuhn: thing is not every C# dev is a python dev
[10:13] Enus Linden: k tao
[10:13] Tao Takashi: I am more focusing on the lib for now
[10:13] Enus Linden: one question though
[10:13] Tao Takashi: Sai: maybe we can provide some web service for testing
[10:13] Enus Linden: pyogp.interop (formerly testharness), needs a buildout to grab crs code does it not?
[10:13] Tao Takashi: but then I wonder if unit tests are really working out.. but maybe they are, might need some experimentation
[10:13] Enus Linden: src*
[10:14] Tao Takashi: Enus: For now it's one buildout which contains lib and interop
[10:14] Saijanai Kuhn: ah, OK. just type in avatar, password, desination sim?
[10:14] Enus Linden: ah
[10:14] Tao Takashi: Sai: maybe also which tests to run
[10:14] Enus Linden: sry
[10:14] Saijanai Kuhn: simple check box sure
[10:15] Enus Linden: so Tao, how should tests in pyogp.interop get run?
[10:15] Tao Takashi: there is a bin/testharness
[10:15] Tao Takashi: which should look for unit tests in interop
[10:16] Tao Takashi: but it wouldn't do configuration
[10:16] Locklainn Linden: in the test folder
[10:16] Tao Takashi: so I wonder how this might work, maybe you want some little app which get some config file as input
[10:16] Enus Linden: so bin/test runs ...? lib unit/doc tests
[10:16] Tao Takashi: bin/test runs pyogp.lib.base tests
[10:16] Enus Linden: bin/testharness runs interop stuff
[10:16] Tao Takashi: bin/testharness runs tests in pyogp.interop
[10:16] Enus Linden: awesome
[10:16] Enus Linden: ty
[10:16] Tao Takashi: in tests/ yes
[10:16] Enus Linden: needs to read the wikis more
[10:17] Bartholomew Kleiber: a bit like best of both worlds, cool
[10:17] Tao Takashi: maybe it's not on there yet. so maybe put it on there then :)
[10:17] Enus Linden: i get bugs in the interop tests. one i don't mind (i can fix it)
[10:17] Enus Linden: the other is a config thing
[10:17] Enus Linden: Module: pyogp.interop.tests.test_r
Traceback (most recent call last):
File "c:\svn\pyogp\libdev\src\py
ez_avatar.py", line 4, in <module>
from indra.ipc import llsdhttp
ImportError: No module named ipc
[10:17] Enus Linden: crap
[10:18] Enus Linden: stupid windows
[10:18] Tao Takashi: yes, this needs to become an egg
[10:18] Enus Linden: hate windows (but run sl on it so have to paste from windfows)
[10:18] Tao Takashi: it's no egg yet as I only eggified the lib requirements for now
[10:18] Enus Linden: can i assign that task to you tao?
[10:18] Enus Linden: or is that a LL thing to do?
[10:18] Tao Takashi: yes but the best thing would be if some indra maintainer could be with me in that boat :)
[10:18] Tao Takashi: so we can basically put the setup.py in the original source code
[10:19] Tao Takashi: right now it's doubled
[10:19] Tao Takashi: and I added a test and buildout for indra
[10:19] Enus Linden: i'm not clear on what needs to happen
[10:19] Enus Linden: i'll jirafy and we can take it to irc...
[10:19] Tao Takashi: [2]
[10:20] Tao Takashi: some person responsible for indra.* should contact me and get the setup.py so he can put it into the original source code
[10:20] Tao Takashi: then I can check this out and make an egg out of it
[10:20] Enus Linden: ah, i see
[10:20] Tao Takashi: well, I also changed the structure a bit so they can be individual eggs
[10:21] Enus Linden: is there a short term work around while we wait for that to happen?
[10:21] Tao Takashi: yes, we double the code as done right now
[10:21] Tao Takashi: so I will simply do the same for indra.ipc
[10:21] Tao Takashi: I just put it into svn for convenience
[10:21] Tao Takashi: so I will do that soon
[10:21] Tao Takashi: do you make a ticket for the merging of those?
[10:22] Tao Takashi: we shouldn't forget about this
[10:22] Enus Linden: yeah tao, jira runs LL. : ) we should track this need there
[10:23] Enus Linden: k, i need to run
[10:23] Locklainn Linden: yea, talk to you all later
[10:23] Enus Linden: i'll be on irc for the bulk of the day
[10:23] Enus Linden: thanks all
[10:23] Bartholomew Kleiber: ok
[10:23] Tao Takashi: and I have to head home
[10:24] Tao Takashi: good meeting :)
[10:24] Locklainn Linden: yep, for sure
[10:24] Locklainn Linden: bye everyone
[10:24] Bartholomew Kleiber: bye all
[10:24] Enus Linden: tao, i do want to take up future dev strategy and roadmap next time around
[10:24] Saijanai Kuhn: I'll post the chat log

Pyogp/Chat Logs/Daily Meeting/28 jul 2008

Navigation menu

Search